Model comparison
Codestral vs Mistral Medium
Mistral Medium is the stronger model overall, scoring 36.3 to 30.6 on the Noometry Index. Codestral costs 6.7× less per token, which makes it the better buy when Mistral Medium's lead doesn't matter for your workload.
Last verified . 2 shared benchmarks.
Mistral AISummary
- They share 2 benchmarks with published results for both. Codestral scores higher in 0 categories and Mistral Medium in 2 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in coding, where Mistral Medium leads 34.2 to 27.3.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 32.5% for Codestral and 50% for Mistral Medium.
- Codestral is cheaper at $0.30 / $0.90 per million input/output tokens, against $1.50 / $7.50 for Mistral Medium.
- Mistral Medium accepts more context: 262K tokens versus 256K.
- Mistral Medium has downloadable open weights; the other is API-only.
Side by side
| Codestral | Mistral Medium | |
|---|---|---|
| Provider | Mistral AI | Mistral AI |
| Noometry Index | 30.6 | 36.3 |
| Released | 2024-05-29 | 2023-12-11 |
| Weights | Proprietary | Open |
| Context window | 256K | 262K |
| Max output | 8K | 262K |
| Input $ / M tokens | $0.30 | $1.50 |
| Output $ / M tokens | $0.90 | $7.50 |
| Results tracked | 7 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral Medium leads
Codestral: 27.3 (#321), Mistral Medium: 34.2 (#243)
| Benchmark | Codestral | Mistral Medium |
|---|---|---|
| ALE-Bench | 137.78 | 763.98 |
| FrontierCode | — | 8% |
| Aider Polyglot | 11.1% | — |
| SciCode | — | 40.2% |
| WeirdML | — | 43.7% |
| BigCodeBench Instruct | 41.8% | — |
| LMArena Coding | — | 1434 |
| BigCodeBench Complete | 52.5% | — |
| HumanEval+ | 73.8% | — |
| MBPP+ | 61.9% | — |
Agentic & Tool Use Not comparable
Codestral: —, Mistral Medium: 28.3 (#90)
| Benchmark | Codestral | Mistral Medium |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 37.7% |
Reasoning Mistral Medium leads
Codestral: 19.8 (#251), Mistral Medium: 24.0 (#167)
| Benchmark | Codestral | Mistral Medium |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | 50% |
| CritPt | — | 0% |
| LMArena Hard Prompts | — | 1426 |
| DTBench | — | 75.5% |
| LMCA | — | 26.1% |
| Surface Evolver Bench | — | 26.9% |
Math Not comparable
Codestral: —, Mistral Medium: 28.1 (#245)
| Benchmark | Codestral | Mistral Medium |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 32.2% |
| ProofBench | — | 9% |
| LMArena Math | — | 1408 |
| MATH Level 5 | — | 81.6% |
| FrontierMath (Feb 2025 set) | — | 0.3% |
Knowledge Not comparable
Codestral: —, Mistral Medium: 25.0 (#265)
| Benchmark | Codestral | Mistral Medium |
|---|---|---|
| GPQA Diamond | — | 59.5% |
| Humanity's Last Exam | — | 4.5% |
| Vectara Hallucination Rate | — | 22.7% |
| LMArena Expert | — | 1408 |
Multimodal Not comparable
Codestral: —, Mistral Medium: 35.3 (#88)
| Benchmark | Codestral | Mistral Medium |
|---|---|---|
| LMArena Vision | — | 1172 |
Multilingual Not comparable
Codestral: —, Mistral Medium: 52.1 (#91)
| Benchmark | Codestral | Mistral Medium |
|---|---|---|
| LMArena Non-English | — | 1408 |
| LMArena Chinese | — | 1447 |
| LMArena French | — | 1459 |
| LMArena German | — | 1432 |
| LMArena Japanese | — | 1378 |
| LMArena Korean | — | 1380 |
| LMArena Russian | — | 1411 |
| LMArena Spanish | — | 1433 |
Instruction Following Not comparable
Codestral: —, Mistral Medium: 73.7 (#116)
| Benchmark | Codestral | Mistral Medium |
|---|---|---|
| LMArena Instruction Following | — | 1398 |
Long Context Not comparable
Codestral: —, Mistral Medium: 42.9 (#114)
| Benchmark | Codestral | Mistral Medium |
|---|---|---|
| LMArena Longer Query | — | 1406 |
Writing & Preference Not comparable
Codestral: —, Mistral Medium: 60.0 (#103)
| Benchmark | Codestral | Mistral Medium |
|---|---|---|
| LMArena Text | — | 1424 |
| LMArena Creative Writing | — | 1391 |
| Short-Story Creative Writing | — | 77.3% |
| LMArena Multi-Turn | — | 1418 |
Frequently asked questions
Is Codestral better than Mistral Medium?
Mistral Medium is the stronger model overall, scoring 36.3 to 30.6 on the Noometry Index. Codestral costs 6.7× less per token, which makes it the better buy when Mistral Medium's lead doesn't matter for your workload.
Which is cheaper, Codestral or Mistral Medium?
Codestral is cheaper. It lists at $0.30 per million input tokens and $0.90 per million output tokens; Mistral Medium lists at $1.50 and $7.50.
Is Codestral or Mistral Medium better for coding?
Mistral Medium scores higher on coding benchmarks: 34.2 versus 27.3 in the Noometry coding category.
Which has the bigger context window?
Mistral Medium does, with 262K tokens against 256K.
How many benchmarks do Codestral and Mistral Medium share?
2 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and Mistral Medium has 36.